Re: summary of locks overflow

It doesn't sound like update stats, caused the overflow. It sounds like it stopped it from happening. If the stats weren't updated the database wouldn't use the indexes and would scan the table. This is especially bad on a join with repeatable read isolation. The database might choose a hash join in memory and lock all the rows in both tables. Scott Kolaya DBA Fleet Services John Carlson <carlson1@bellsouth.net> wrote in article <34FC6C3D.1854@bellsouth.net>... > hkchan@idt.com wrote: > > > > hi, > > > > . . . snip . . . > > > > > ultimately it was not our onconfig that's causing the locks overflow, > > we'd allocated 200000 to it, but in forgetting to run update stats. > > at least that's the conclusion that we came to. anyone else hv any > > ideas on this? thanks. > > Update stats causing a lock overflow? That's interesting. > Are you sure that it wasn't an application that locked a few too many > rows? > > > John Carlson > Informix DBA > WH Smith, Inc. >